Measurement errors are broadly classified by how they behave when a measurement is repeated. The key distinction is whether the error stays the same each time (systematic) or varies unpredictably (random).

The scenario describes measuring the same quantity several times and getting readings that vary slightly from one trial to the next. This scatter is the signature of a random error, which is the correct answer. Random errors arise from small, uncontrollable fluctuations such as instrument electrical noise, tiny changes in temperature or supply, observer inconsistency, and vibration. Because they are equally likely to push readings up or down, they can be reduced by taking many readings and averaging, and they are characterised statistically (e.g. by standard deviation).

Why the other options are wrong:

• Systematic error produces a consistent offset in one direction (readings always too high or always too low by roughly the same amount); it does not by itself cause the reading-to-reading scatter described here, and averaging does not remove it.
• Calibration error is a specific kind of systematic error (an instrument reading consistently off due to poor calibration), so it too shows a fixed bias rather than random fluctuation.
• Parallax error comes from viewing a scale or pointer at the wrong angle and is essentially systematic in nature; it would not naturally produce slightly different values on every repeated reading.